搜索内容 搜索用户

# 14 麦片

050

23主题 2最佳答案
• 关注者： 1

 matlab中有一个函数是autocorr 我查看它的帮助 AUTOCORR Compute or plot sample auto-correlation function.     Compute or plot the sample auto-correlation function (ACF) of a univariate,     stochastic time series. When called with no output arguments, AUTOCORR     displays the ACF sequence with confidence bounds.     [ACF, Lags, Bounds] = autocorr(Series)     [ACF, Lags, Bounds] = autocorr(Series , nLags , M , nSTDs)     Optional Inputs: nLags , M , nSTDs   Inputs:     Series - Vector of observations of a univariate time series for which the       sample ACF is computed or plotted. The last row of Series contains the       most recent observation of the stochastic sequence.   Optional Inputs:     nLags - Positive, scalar integer indicating the number of lags of the ACF       to compute. If empty or missing, the default is to compute the ACF at       lags 0,1,2, ... T = minimum[20 , length(Series)-1]. Since an ACF is       symmetric about zero lag, negative lags are ignored.     M - Non-negative integer scalar indicating the number of lags beyond which       the theoretical ACF is deemed to have died out. Under the hypothesis that       the underlying Series is really an MA(M) process, the large-lag standard       error is computed (via Bartlett's approximation) for lags > M as an       indication of whether the ACF is effectively zero beyond lag M. On the       assumption that the ACF is zero beyond lag M, Bartlett's approximation       is used to compute the standard deviation of the ACF for lags > M. If M       is empty or missing, the default is M = 0, in which case Series is       assumed to be Gaussian white noise. If Series is a Gaussian white noise       process of length N, the standard error will be approximately 1/sqrt(N).       M must be less than nLags.     nSTDs - Positive scalar indicating the number of standard deviations of the       sample ACF estimation error to compute assuming the theoretical ACF of       Series is zero beyond lag M. When M = 0 and Series is a Gaussian white       noise process of length N, specifying nSTDs will result in confidence       bounds at +/-(nSTDs/sqrt(N)). If empty or missing, default is nSTDs = 2       (i.e., approximate 95% confidence interval).   Outputs:     ACF - Sample auto-correlation function of Series. ACF is a vector of       length nLags + 1 corresponding to lags 0,1,2,...,nLags. The first       element of ACF is unity (i.e., ACF(1) = 1 = lag 0 correlation).     Lags - Vector of lags corresponding to ACF (0,1,2,...,nLags).     Bounds - Two element vector indicating the approximate upper and lower       confidence bounds assuming that Series is an MA(M) process. Note that       Bounds is approximate for lags > M only. 我始终弄不明白M 到底有什么用处，请各位高手帮帮忙？

# 14 麦片

050

23主题 2最佳答案
• 关注者： 1
楼主| 发表于 2007-12-26 10:05:00 | 显示全部楼层

## 自己的问题自己回答

 我原来误解了这个函数的作用 举一个例子来说把,autocorr(sm,[],2,2) autocorr()函数是时间序列自相关函数 sm : 一个时间序列数据 []: 表示计算这个时间序列数据的自相关函数的延迟. 2: 表示自相关函数在>2的所有延迟的自相关系数看作为0 2: 表示在什么范围内时间序列的数据被看作是0

# 14 麦片

050

23主题 2最佳答案
• 关注者： 1
楼主| 发表于 2007-12-26 10:06:22 | 显示全部楼层
 希望自己抓紧努力,争取学有建树

# 5 麦片

050

1主题 0最佳答案

 (⊙o⊙)…还是不懂

# 5 麦片

050

4主题 0最佳答案
• 关注者： 1

 高手、达人们快来解释一下啊，楼主的解答已经不错了，但是详细的还不是太明白。希望高手们能详细解答下。。。

# 11 麦片

050

2主题 0最佳答案

 同问这个M(2)代表什么意思

# 5 麦片

050

0主题 0最佳答案

 tiandao091 发表于 2013-5-17 16:16 同问这个M(2)代表什么意思 滞后时长，如果是默认的就是以高斯白噪声。

# 11 麦片

050

1主题 0最佳答案

 soga  我还要再看看

# 11 麦片

050

1主题 0最佳答案

 M--非负整数，表示在多大延迟时理论ACF为0.autocorr假设序列为MA(M)，并且使用Bartlett估计方法来计算大于M的延迟的标准误差。如果M=[]或缺省，则为0，函数假设序列为高斯白噪声。

# 5 麦片

050

0主题 0最佳答案

 那autocorr只能算20个点吗，不能增加步长吗

 您需要登录后才可以回帖 登录 | 注册 本版积分规则 回帖后跳转到最后一页